Energy Efficient Event Localization and Classification for Nano IoT

TL;DR

This paper introduces an energy-efficient nano IoT framework where nanosensors detect events and communicate their location and type using single pulses, enabling accurate localization and classification with minimal energy consumption.

Contribution

The paper proposes a novel nano IoT event monitoring method that uses spectral encoding and direction estimation for energy-efficient localization and classification.

Findings

Achieves 1.58-degree localization accuracy at 1 meter distance.

Classifies event types with 98.8% accuracy using minimal energy.

Operates with atto Joule energy pulses.

Abstract

Advancements in nanotechnology promises new capabilities for Internet of Things (IoT) to monitor extremely fine-grained events by deploying sensors as small as a few hundred nanometers. Researchers predict that such tiny sensors can transmit wireless data using graphene-based nano-antenna radiating in the terahertz band (0.1-10 THz). Powering such wireless communications with nanoscale energy supply, however, is a major challenge to overcome. In this paper, we propose an energy efficient event monitoring framework for nano IoT that enables nanosensors to update a remote base station about the location and type of the detected event using only a single short pulse. Nanosensors encode different events using different center frequencies with non overlapping half power bandwidth over the entire terahertz band. Using uniform linear array (ULA) antenna, the base station localizes the events by estimating the direction of arrival of the pulse and classifies them from the center frequency estimated by spectral centroid of the received signal. Simulation results confirm that, from a distance of 1 meter, a 6th derivative Gaussian pulse consuming only 1 atto Joule can achieve localization and classification accuracies of 1.58 degree and 98.8%, respectively.